Upgraded Desktop: New issue with the Graphics Driver stopping and crashing

Duff0712

Reputable
Dec 20, 2015
7
0
4,510
First off I am not extremely tech savvy (I know how to build a computer etc, but still have a lot to learn), so please keep with me. I learn fast! I upgraded some older computer parts a week ago and I have been having issues when gaming.

Most games I have played tend to freeze up no more than 30 minutes of play. Sometimes only 3 minutes. The screen usually turns "windows" blue or a tan color and the game will pop back up, with a message in the bottom right saying the "Application has been blocked from accessing Graphics hardware". I researched a ton and tried several fixes to no avail. Now when playing I usually receive the messages "Display Drivers have stopped working and have recovered successfully" twice then my computer fully freezes with a sound loop.

This has happened playing League of Legends and the Blade n Soul beta. I haven't noticed it happen in Star Wars Battlefront or Fallout 4 yet, but I haven't been playing them that much, since I've trying to fix the issue.

My Specs (* means upgraded part):
- Windows 10 home 64bit
- Intel Core i5-3570K (BX80637I5350K)
- MSI 1155 Motherboard (Z77A-GD55)
- Cooler Master Extreme 700w (RS700-PCAAE3-US)
- *EVGA Geforce GTX 970 ACX 2.0+ (04G-P4-3975-KR)
- *850 EVO 500GB SSD (MZ-75E500B/AM)
- *G.SKILL Trident X 16GB (2x8GB) DDR3 1600 (F3-1600C7D-16GTX)

Things I have tried so far:

- Changing power management mode to "Prefer maximum performance"
- Uninstalling and rolling back driver with DDU. I have only tried going back to 359.00 (from 359.06).
- Running an "all-in-one" stress-test (Furmark?). Nothing crashing during.
- Turning V-sync off
- Probably more, have forgotten though.

I have installed several programs like GPU-Z, MSI Afterburner and Speccy to keep an eye on my system, but not 100% sure what everything means. My temperatures seem ok (GPU never exceeding 69C). I have no idea what is causing the issue...

One other thing that might be worth mentioning. While my computer was idling my GPU seemed to be using more resources then after I had started a game and exited it in fear of a computer crash.

GPU-Z idle:
- Core Clock: 1189MHz
- Memory Clock: 1752MHz
- Temp: 56C
- Fan Speed: 0%
- Power Consumption: 28%
- VDDC: 1.0310V

Idle after I started BnS beta and exiting the game after 3 min because of the Driver stopped issues:
- Core Clock: 135MHz
- Memory Clock: 162MHz
- Temp: 33C
- Fan Speed: 0%
- Power Consumption: 7.8%
- VDDC: .8680V

Everything will jump back up for a couple seconds every few minutes, but everything has significantly dropped since then. Those are the biggest difference I've seen in GPU-Z. The only other thing I've seen is in PerfCap Reason and the only thing that has shown up is VRel.

If anyone has any ideas on what it may be or what to try please let me know. This is super frustrating.

Was testing out a couple games after rerolling to 359.00. Played Star Wars Battlefront for 2+ hours with no crashes or frame loss (90-100) whatsoever. Played the BnS Beta for 2 minutes and got the driver error again. I have a couple screenshots from GPU-Z

This one is from playing Star Wars Battlefront.
2YIFhzT.gif


This one is from playing BnS for 2 minutes. (The game was started around the middle, scrolling over the GPU load it hit 100 at the peak. Not sure why it would do that... My guess is it happens when I play League of Legends too)
U21lBaU.gif

 

Duff0712

Reputable
Dec 20, 2015
7
0
4,510


If it is overclocking, I haven't done it on purpose. I haven't messed with the card at all.
 

Duff0712

Reputable
Dec 20, 2015
7
0
4,510


I will try that. I already have afterburner, but I will look up how to do that and test it.

update: So I went into MSI Afterburner and changed the Memory Clock MHz to -25. Seems like that fixed my BnS problem! I was able to play the game for an hour and no driver error or crashes. My fps also improved. Here is what GPU-Z read
9YceVXP.gif


After a bit I tried Star Wars Battlefron as well. I wasn't having problems with it before, but wanted to see how it would do. The GPU-Z read almost exactly the same, however the GPU load was at a constant 99%. My framerate also went from a 90-100 down to 30-35. I tried resetting the afterburner settings to see it if would fix it, but it didn't. I had to restart my computer and although it fixed the framerate on Battlefront. Now GPU-Z is reading this again on idle. Causing my BnS to crash every couple minutes...
3sJZxu9.gif
 

ThatGuyTyping

Reputable
Dec 15, 2015
48
0
4,540
I'd keep tweaking your memory settings. Your core clock is running at stock speeds. I'd try to bring you core clock up I've seen a number of people add ~ 100 to the core. Then bring the memory down until you are getting a good result in the games you play.
 

Duff0712

Reputable
Dec 20, 2015
7
0
4,510


Thanks I will give that a try and update the thread after. With afterburner, will I need to reactivate the profile I create every time I restart my computer?
 

ThatGuyTyping

Reputable
Dec 15, 2015
48
0
4,540


no, at the bottom under profile you can check the "apply overclocking at system start up". Also in settings check "start with windows"



 

Duff0712

Reputable
Dec 20, 2015
7
0
4,510



Thanks! I appreciate you helping me out :)

I haven't tested changing the Core and Memory clock a lot yet, but I found something interesting.

As a recap, yesterday when I dropped the memory clock MHz -25 in afterburner, my core clock read 539MHz and my memory clock read 1502MHz. BnS Beta played great, but I took a huge FPS drop in Battlefront. Also, after trying to reset afterburner settings, neither clocks would return to "normal" until I restarted my computer.

Today I dropped memory clock the exact same as yesterday, -25, and the core clock stayed the same at 1379, but memory clock only dropped from 1752 to 1734~. Battlefront played fine, but BnS crashed in 30 seconds of play (I haven't played with the numbers too much yet).

I find it weird that my core MHz aren't changing like they were yesterday.
 

Duff0712

Reputable
Dec 20, 2015
7
0
4,510
Sorry for the bump, but I thought I would give an update.

I haven't changed the core speeds yet as I am still unsure of implementing them or how much I should change them.

However, after the nvidia driver update on the 21st I decided to try those drivers. I did another clean install with DDU and decided not to install the 3D options and the audio option. After that I had no issues whatsoever. The BnS beta had ended so I couldn't test that, but I could play any of my games flawlessly.

It worked great for a week and now I am having issues again. My buddy got me Rocket League as a gift and I played that with him for a solid 3 hours the day of (which is super fun haha). The next day I played with him again for about an hour and my computer "crashed". My screen turned pink and the last sound made in game was in an audio loop. I could however still hear my friends talking in discord. I tried to alt/tab and the audio loop stopped but no change in the screen was made. After another alt/tab or alt/cntrl/delete I couldn't hear any sound after that. The only way to fix it was a hard reset. I should also mention that after one of the pink screen crashes I let it sit for 3-5 minutes and the screen turned black after (no mention from the monitor of a loss in signal though).

I changed my fan settings in MSI to see if it was my temp, but even after changing the fans and rarely going above 50C (never broke 70C before changing), it would still happen after like 30 minutes of play. I have not tested any other game yet, but this is getting frustrating.

I will see if I can downclock my card and see if that actually helps (although any guidance on that would be helpful as I don't know exactly what numbers to use). If it does help I will probably RMA it through EVGA or Amazon (where I bought it) since I shouldn't need to downclock it. If it doesn't fix it then I would assume it isn't the card itself and I figure I might try a clean windows 10 install.

This sucks.